Re: Open bugs in 21.5

Hi,

>>>>> In <[email protected]>
>>>>>	Frank Schmitt <[email protected]> wrote:

> The patch is not in CVS, I applied it to my personal copy. As far as I
> understood Katsumi, the patch fixes some issues for Japanese users, but
> there are still many problems, he said something like "It's okay to
> revert the patch since some more problems don't count"

The reverse patch to withdraw my bogus changes is available in:

http://list-archive.xemacs.org/xemacs-beta/200209/msg00454.html

Please apply it.  And return XEmacs sane, please.

XEmacs is my favorite, and I am apprehending the popularity of
XEmacs is lost in Japan.  Actually, no Japanese people use 21.5
beta since April 28.  Do you know the xemacs-beta-ja and xemacs-
users-ja lists are quite deserted?  By the following reasons
that I can remember, we cannot use XEmacs 21.5 beta:

>>>>> In <[email protected]>
>>>>>	Katsumi Yamaoka <[email protected]> wrote:

> There's a bug released for a long time.  XEmacs 21.4.9 returns t,
> but XEmacs 21.5-b9 returns nil for the following expression.

(let ((str (char-to-string (make-char 'japanese-jisx0208 36 34))))
  (string-equal (decode-coding-string
		 (encode-coding-string str 'euc-japan)
		 'undecided)
		str))

> In other words, XEmacs 21.5-b9 doesn't do the automatic decoding
> for Japanese text.  It doesn't work not only for euc-japan but
> also for shift_jis etc.  Probably it relates to this, find-file
> treats Japanese text as raw-text.


>>>>> In <[email protected]>
>>>>>	Katsumi Yamaoka <[email protected]> wrote:

> 2. Since `language-unicode-precedence-list' does not have the
>    default value, XEmacs decodes a unicode encoded Japanese text
>    with Chinese characters:

(char-charset
 (string-to-char
  (decode-coding-string
   (encode-coding-string
    (char-to-string (make-char 'japanese-jisx0208 36 34))
    'utf-8)
   'utf-8)))
 => chinese-gb2312

> 3. The Japanese category table has not been set, so the regexp
>    "\\cj" does not match Japanese text:

(string-match "\\cj"
	      (char-to-string (make-char 'japanese-jisx0208
					 36 34)))
 => nil

> 5. Converting hiragana string to katakana string, or katakana to
>    hiragana doesn't work.  It is because the character code
>    property has not been set for those Japanese characters.

(let* ((hiragana (char-to-string (make-char 'japanese-jisx0208
					    36 34)))
       (katakana (japanese-katakana hiragana)))
  (string-equal hiragana katakana))
 => t
-- 
Katsumi Yamaoka <[email protected]>